Impact of SLC20A1 on the Wnt/Β‑Catenin Signaling Pathway in Somatotroph Adenomas

3276 Molecular Medicine REPORTS 20: 3276-3284, 2019

Impact of SLC20A1 on the Wnt/β‑catenin signaling pathway in somatotroph adenomas

JIANHUA LI1,2, WEI DONG3, ZHENYE LI4, HONGYUN WANG1, HUA GAO1 and YAZHUO ZHANG1,4

1Key Laboratory of Central Nervous System Injury Research, Beijing Neurosurgical Institute, Capital Medical University, Beijing 100070; 2Department of Neurosurgery, Binzhou People's Hospital, Binzhou, Shandong 256610; 3Department of Neurosurgery, Tangshan People's Hospital, Tangshan, Hebei 063001; 4Beijing Tiantan Hospital, Capital Medical University, Beijing 100070, P.R. China

Received February 5, 2019; Accepted July 12, 2019

DOI: 10.3892/mmr.2019.10555

Abstract. Studies have revealed that genetic and functional was 86.4±5.7, 75.6±4.4 and 67.5±3.8%, respectively (P<0.05). aberrations of oncogenes, tumor‑suppressor genes, signaling ELISA analysis demonstrated the growth hormone (GH) pathways and receptors are among the most prominent events levels in the Sh‑B‑SLC20A1 and Sh‑C‑SLC20A1 groups to be in pituitary tumorigenesis, and a potent biomarker would 34.7±10.4 and 54.6±14.4%, respectively, of that of control GH3 be helpful for early diagnosis, subsequent treatment and cells (P<0.05). The transmembrane invasion assay revealed that disease control. The present study investigated the expression knocking down SLC20A1 significantly suppressed cell invasion signatures of solute carrier family 20 member 1, also known in the Sh‑B‑SLC20A1 and Sh‑C‑SLC20A1 groups. RT‑qPCR as phosphate transporter 1 (SLC20A1) and the Wnt/β‑catenin and western blotting demonstrated that Sh‑B‑SLC20A1 and signaling pathway in 52 patients with somatotroph adenomas. Sh‑C‑SLC20A1 evidently increased the levels of Wif1 and According to immunohistochemistry analysis, the H‑score secreted frizzled‑related protein 4. The present data suggested of SLC20A1 was 222.6±15.2 in invasive tumor samples and that SLC20A1 levels are positively associated with tumor 144.5±30.4 in non‑invasive tumor samples (P<0.01), while the size, invasive behavior and tumor recurrence in somatotroph H‑scores of β‑catenin were 210.1±21.4 and 134.9±32.7, respec- adenomas. Furthermore, SLC20A1 may be associated with the tively (P<0.05). The H‑scores of Wnt inhibitory factor 1 (Wif1) activation of the Wnt/β‑catenin signaling pathway. exhibited the opposite trend, with scores of 134.5±22.7 and 253.6±14.8, respectively (P<0.01). The H‑scores of SLC20A1 Introduction were negatively associated with those of Wif1 in somatotroph adenomas (correlation coefficient r=‑0.367). The mean Somatotroph adenomas cause disfiguring growths and progression‑free survival in the low SLC20A1 group was longer medical complications, including cardiovascular and pulmo- than that in the group with high SLC20A1 H‑scores (P=0.024). nary complications, and increased mortality rates due to Reverse transcription‑quantitative PCR (RT‑qPCR) and western malignancy (1). Acromegaly is mainly caused by somato- blotting confirmed the interference efficiency of the segments troph adenomas that secrete excess growth hormone (GH) short hairpin (Sh)‑B‑SLC20A1 and Sh‑C‑SLC20A1. Cell and insulin‑like growth factor 1 (IGF‑1) (2). Somatotroph proliferation experiments revealed that the cell viability of the adenomas are initially treated surgically, followed by radia- Sh‑B‑SLC20A1 group was 76.3±4.5, 65.7±3.7 and 53.1±3.2% of tion and/or medication, if necessary (3). Large tumor sizes and that of control GH3 cells after 24, 48 and 72 h of transfection, high post‑operative IGF‑1 levels are predictors for incomplete respectively, while the cell viability of the Sh‑C‑SLC20A1 group remission and uncontrolled disease (4). In a complex multistep process, genetic and epigenetic factors, hormonal stimulation, growth factors and their receptors contribute to tumorigenesis and the development of somatotroph adenomas. To date, no high‑frequency somatic genetic alterations have been identi- Correspondence to: Dr Hua Gao or Dr Yazhuo Zhang, Key fied in somatotroph adenomas, and the somatic landscape Laboratory of Central Nervous System Injury Research, Beijing shows Ca+2 and the ATP signaling pathways to be involved in Neurosurgical Institute, Capital Medical University, 119 Southwest pituitary tumorigenesis (5). However, there is a lack of clinical 4th Ring, Beijing 100070, P.R. China evidence of phenotype‑genotype correlations between the E‑mail: [email protected] E‑mail: [email protected] α subunit of stimulatory G‑protein (GNAS) in patients with mutated and non‑mutated somatotroph adenomas (6). Key words: somatotroph adenomas, solute carrier family 20 Inorganic phosphate (Pi) enters the cells via Na/Pi cotrans- (phosphate transporter), member 1, Wnt/β‑catenin signaling pathway, porters, which comprise solute carrier (SLC)20 and SLC34 in Wnt inhibitory factor 1, invasion mammals (7). The SLC20 family consists of two members, namely SLC20A1 [also known as phosphate transporter 1 (PiT‑1)] and SLC20A2 (PiT2). SLC20A1 encodes sodium‑dependent LI et al: SLC20A1 IN SOMATOTROPH ADENOMAS 3277 phosphate transporter 1, whose gene is located on chromosome described method (15). The following primary antibodies were 2q11‑14, a log 10 odds peak score region (8). Overexpression of used: Anti‑SLC20A1 (cat. no. ab237527; 1:1,000; Abcam), SLC20A1 increases cell proliferation and colony formation in anti‑β‑catenin (cat. no. ab32572; 1:2,000; Abcam), anti‑Wif1 murine fibroblastic NIH3T3 and pre‑osteoblastic MC3T3‑E1 (cat. no. ab155101; 1:500; Abcam) and anti‑Ki 67 (cat. no. ACK02; cells. Furthermore, non‑proliferating cells exhibit the lowest 1:1,000; Leica Microsystems, Inc.). A secondary antibody SLC20A1 mRNA levels in these cell lines (9). Overexpression (cat. no. sc‑2040; 1:5,000; Santa Cruz Biotechnology, Inc.) was of SLC20A1 leads to faster adhesion and spreading compared used in combination with the Bond Polymer Refine Detection with control NIH3T3 cells (10). High SLC20A1 expression was system (cat. no. DS9800; Leica Microsystems, Inc.). Staining associated with worse survival in 401 patients with estrogen intensity was stratified on a scale of 0‑3 as follows: 0, no staining; receptor‑positive breast tumors compared with low SLC20A1 1, weak staining; 2, moderate staining; and 3, strong staining. The expression (10‑year survival rates, 17.16 vs. 68.45%) (11). The H‑score was obtained by multiplying the staining intensity by a epidermal growth factor receptor inhibitor gefitinib exhibits constant to adjust the mean to the strongest staining, to produce strong inhibition of cell proliferation in glioblastoma cell lines a score in the range of 0‑300. Specifically, H‑score=scale x by regulating SLC20A1 expression (12). In addition, SLC20A1 percentage of strong staining (0‑100%). H‑score=1.0 indicated a modulates erythroid maturation by modulating the activity of weak percentage; 2.0 indicated a moderate percentage; and 3.0 erythroid Kruppel‑like factor in vivo and in vitro (7). indicated a strong percentage. The Wnt signaling pathway has been implicated in the processes of tissue differentiation, proliferation, apoptosis Cell culture, proliferation and migration. GH3 cells were and tumorigenesis (12). The levels of Wnt inhibitory factor 1 purchased from the American Type Culture Collection (ATCC) (Wif1) and secreted frizzled‑related protein 4 (sFRP4) are and were cultured in F12K (ATCC) supplemented with 2.5% lower in invasive nonfunctional pituitary adenomas (NFPAs) fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) and than in noninvasive NFPAs (13). The present study investigated 15% horse serum (Gibco; Thermo Fisher Scientific, Inc.) under the expression signatures of SLC20A1 and the Wnt/β‑catenin 37˚C and 5% CO2. signaling pathway in 52 patients with somatotroph adenomas. The GH3 cells were plated onto 96‑well dishes with Furthermore, it explored the possibility that SLC20A1 could 1x104 cells and 100 µl medium in each well, and incubated be targeted by chemotherapeutic agents or used as a molecular overnight. SLC20A1 gene silencing in the GH3 cells was marker for the diagnosis of patients with cancer via RNA established by transfection of short hairpin RNA (1 µg for interference (RNAi) technology in the GH3 cell line. 1x106 cells; OriGene Technologies, Inc.) or non‑targeting control shRNA (scramble) for 72 h using 5 µl Opti‑MEM Materials and methods (Gibco; Thermo Fisher Scientific, Inc.) and 0.3 µl Lipofectamine® 3000 (Invitrogen; Thermo Fisher Scientific, Patients and tissue specimens. The present study retro- Inc.), according to the manufacturer's instructions. The spectively reviewed 52 patients (mean age, 40.3 years; range sequences of the pGFP‑C‑shLenti plasmid were: Scramble, 16‑62 years) with somatotroph adenomas at Beijing Tiantan TTCTCCGAACGTGTCACGT; A, CACCGAAGTATGACA Hospital affiliated to Capital Medical University, between ATCTGTGGATGCTC; B, GCAATGCTGTGTCTGACC July 2012 and December 2016. The diagnosis of somatotroph TTCACTCAGAG; C, ATAGGAATCCTGTGTCTGAGG adenoma was based on clinical symptoms and histopatho- TAGTATGT; and D, GCTTCTGCTCCACCGAAGTATGAC logical features, per the 2017 World Health Organization AATCT. Then, 20 µl MTS solution was added to each well Classification (14). All patients were diagnosed with sporadic and further incubated for 4 h. The absorbance at 490 nm of somatotroph adenomas. The inclusion criteria for recruitment each well was measured using an ELISA plate reader (Thermo of patients were as follows: i) Serum GH >1 ng/ml and IGF‑1 Fisher Scientific, Inc.). above normal (115‑307 ng/ml), or GH >2 ng/ml; ii) magnetic Cell migration and invasion were measured using fibro- resonance imaging confirmed the existence of the saddle area; nectin‑ and Matrigel‑coated polycarbonate filters, respectively, iii) acromegaly or facial changes; and iv) transmission electron and modified transwell chambers (Corning Inc.). In total, microscopy revealed large secretory granules (400‑600 nm). 5x104 GH3 cells in 200 µl F12K medium were added into Patients presenting familial pituitary adenoma were excluded the upper chambers and 600 µl medium [F12K + 2.5% FBS from the present study. In addition, three normal pituitary (cat. no. 10099‑141; Gibco; Thermo Fisher Scientific, Inc.)+ glands were obtained from a donation program at Tianjin First 5% horse serum (cat. no. sh30074; HyClone; GE Healthcare Central Hospital in China (two male and one female patients; Life Sciences] was added into lower chamber. After 24 h, age, 21‑45). None of the donors had a history of pituitary disease. migrating cells that adhered to the lower membrane were Progression‑free survival (PFS) was defined as the interval fixed in 4% paraformaldehyde (for 30 min at 4˚C) and stained between the date of surgery and the date of tumor recurrence or using hematoxylin (OriGene Technologies, Inc.) for 5 min at tumor regrowth. The protocols were approved by the Internal room temperature. The average number of migrated cells was Review Board of Beijing Tiantan Hospital affiliated to Capital quantified in five randomly selected fields of view under a Medical University and the study was conducted according to ﬂuorescence microscope (Zeiss GmbH; magnification, x100) the principles of the Declaration of Helsinki (no. KY‑2013‑02). Experiments were performed in triplicate. Informed consent was obtained from all patients. ELISA. GH level was detected using an ELISA kit Immunohistochemistry (IHC). Tissue microarray construction (cat. no. 710685; Shanghai Enzyme‑linked Biotechnology Co., and the IHC protocol were performed according to a previously Ltd.) according to the manufacturer's protocol. The absorbance 3278 Molecular Medicine REPORTS 20: 3276-3284, 2019

Table I. Primers used in reverse transcription‑quantitative PCR.

Gene Forward primer (5'‑3') Reverse primer (5'‑3')

SLC20A1 gctttatggtggtgttggca GTGTTGTGctgatgggaagg E‑CAD actttggtgtgggtcaggaa cacatgctcagcgtcttctc N‑CAD agaacagggtggacgtcatt accactgtgactagcccatc MMP‑2 TGcaaccacaaccaactacg tagagctcctggatcccctt MMP‑9 TGGGcaagcagtactctacc gtcttcatgcagaggggagt Snail cgagcagagttgtctaccga ctgctggaaggtgaactcca Vimentin actaatgagtccctggagcg aggtggcgatctcaatgtca VEGF caccaaagccagcacatagg tttaactcaagctgcctcgc GAPDH agtctactggcgtcttcacc ccacgatgccaaagttgtca

SLC20A1, solute carrier family 20 member 1; E‑CAD, E‑cadherin; N‑CAD, N‑cadherin; MMP, matrix metalloproteinase; VEGF, vascular endothelial growth factor.

at 450 nm in each well was measured using an ELISA plate KGaA) at 4˚C overnight, and then incubated 1 h with horse- reader (M200 Pro; Tecan Group, Ltd.). radish peroxidase‑tagged secondary antibodies (cat. no. sc‑2363, 1:5,000; Santa Cruz Biotechnology, Inc.) at room temperature. Reverse transcription‑quantitative PCR (RT‑qPCR) Finally, the membranes were visualized by enhanced chemilu- analysis. Total RNA was extracted from 24 frozen pitu- minescence (cat. no. sc2048, Santa Cruz Biotechnology, Inc.) itary adenoma samples (~10 mg) using the RNeasy® Mini using a Versadoc XL imaging system (Amersham Imager 600; Kit (Qiagen GmbH). Purified total RNA (1 µg) was reverse GE Healthcare Life Sciences), and densitometry was performed transcribed into complementary DNA (cDNA) using a using the ImageQuant TL software (version 7.0; GE Healthcare Revert Aid First‑Strand cDNA Synthesis kit (Thermo Fisher Life Sciences). GAPDH levels were analyzed as a loading control. Scientific, Inc.), based on the manufacturer's protocol (25˚C for 10 min, 37˚C for 2 h, 95˚C for 5 min). The primer details Statistical analysis. The χ2 exact test was used to determine are presented in Table I. GAPDH was used as an endogenous the significance of clinicopathological characteristics and vari- control for normalizing the levels of target genes. RT‑qPCR ables. Pearson's correlation coefficient was used to analyze the was performed in an ABI 7500 Fast Real‑Time PCR System correlation between β‑catenin/Wif1 expression and SLC20A1 (Applied Biosystems; Thermo Fisher Scientific, Inc.) using score. Student's t‑test was applied to examine the differential Platinum® SYBR® Green qPCR SuperMix‑UDG (Invitrogen; expression of SLC20A1, β‑catenin, Ki67 and Wif1 in patient Thermo Fisher Scientific, Inc.). The comparative quantifica- samples. Data are presented as the mean ± SD. One‑way tion cycle (Cq) method was used to calculate the fold‑change ANOVA was used in cell growth, GH release and migration of in differential expression of each gene (2‑ΔΔCq method) (16) GH3 cells experiments. All P‑values were two sided and P<0.05 the thermocycling conditions were the following: Initial was considered to indicate a statistically significant difference. denaturation at 95˚C for 15 followed by 40 cycles of 94˚C for All experiments were repeated three times. Analyses were 15 sec, 55˚C for 30 sec and 74˚C for 34 sec. performed using SPSS (vesion 19.0; IBM Corp.).

SDS‑PAGE and western blot analyses. A total of 10 mg specimen Results was lysed to extract total protein in TNE buffer (50 mM Tris‑HCl, pH 7.4, 150 mM NaCl and 1 mM EDTA; Sigma‑Aldrich; Merck Clinicopathological features. There were 28 males and KGaA) containing 1% Nonidet P‑40 (Calbiochem; Merck 24 females in the present study (Table SI) with a mean age of KGaA) with protease and phosphatase inhibitor cocktails (Roche 40.3 years (age range, 17‑62 years). The median tumor volume Applied Science). The total protein extracted from somato- was 6.25 cm3 (range, 2.24‑18.7 cm3). The average level of GH troph adenomas or normal pituitary glands was centrifuged at was 22.83±7.31 ng/ml, and ranged from 7.9 to 76.2 ng/ml. 12,000 x g for 30 min at 4˚C, and the protein concentration was The number of total resections was 31/52 (59.6%), while the determined using a bicinchoninic acid protein assay kit (Pierce; number of partial resections was 21/52 (40.4%). The number Thermo Fisher Scientific, Inc.). For western blot analysis, 40 µg of recurrent cases was 11/52 (21.2%). Tissues from 52 patients protein per lane was separated electrophoretically using 4‑12% were stained with standard hematoxylin and eosin staining, Bis‑Tris SDS‑PAGE gels and blotted onto polyvinylidene fuo- and IHC was performed with antibodies against SLC20A1, ride membranes. Each membrane was blocked for 1 h with 5% β‑catenin, sFRP4, Ki67 and Wif1. According to the Knosp milk in TBS at room temperature, incubated with antibodies classification, the 52 patients were classified into an invasive against SLC20A1 (cat. no. ab237527, 1:2,000; Abcam), Wif1 group (15 patients) and a non‑invasive group (37 patients). The (cat. no. ab155101, 1:1,000; Abcam), β‑catenin (cat. no. ab32527, patients in the invasive group had a greater degree of head- 1:5,000; Abcam), sFRP4 (cat. no. ab154167, 1:2,000; Abcam) ache and visual deficits than the patients in the non‑invasive and GAPDH (cat. no. G5262, 1:8,000; Sigma‑Aldrich; Merck group. The preoperative serum GH level was 36.7±11.9 ng/ml LI et al: SLC20A1 IN SOMATOTROPH ADENOMAS 3279 in the invasive group and 17.2±6.3 ng/ml in the non‑invasive Table II. Univariate analyses of clinicopathological correlates group (P<0.05). Follow‑up data revealed that total resection in 52 patients with somatotroph adenomas. was carried out in 5/15 (33.3%) cases in the invasive group and 26/37 (70.3%) cases in the non‑invasive group (χ2=4.611, Univariate P=0.032). Furthermore, recurrence in the invasive group invasiveness analysis occurred in 6/15 (40%) cases and in 5/37 (13.5%) cases in the ‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑ ‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑ non‑invasive group (χ2=4.489, P=0.034) (Table II). Variable Yes (n=15) No (n=37) χ² P‑value

SLC20A1, Wif1, β‑catenin and Ki67 expression signatures Age, years 2.342 0.126 in 52 somatotroph adenomas. IHC was used to evaluate the ≤40.3 10 16 levels of SLC20A1, Wif1, β‑catenin and Ki67 in the 52 somato- >40.3 5 21 troph adenoma samples (Fig. 1A). The H‑score of SLC20A1 Sex 0.843 0.358 was 222.6±15.2 in the invasive group and 144.5±30.4 in the Male 9 17 non‑invasive group (P<0.01), while the H‑scores of β‑catenin Female 6 20 were 210.1±21.4 and 134.9±32.7 in these two groups, respec- Tumor size, cm3 5.996 0.014 tively (P<0.05). The H‑scores of Wif1 presented the opposite trend compared with β‑catenin or SLC20A1, with values of ≤6.25 12 14 134.5±22.7 and 253.6±14.8 in the invasive and non‑invasive >6.25 3 23 groups, respectively (P<0.01). The Ki67 index was 6.3±2.2 GH level, ng/ml 36.7±11.9 17.2±6.3 - <0.05 in the invasive group and 1.9±1.3 in the non‑invasive group. Resection 4.611 0.032 Western blotting also confirmed the same tendency, as depicted Total 5 26 in Fig. 1B. In the present study, a high‑level case was defined Partial 10 11 as having a H‑score higher than the median value. Recurrence 4.489 0.034 According to the H‑score median of SLC20A1, Wif1 Yes 6 5 and β‑catenin, 52 samples were divided into high and low H‑score groups. The tumor volume in patients with high No 9 37 SLC20A1 H‑scores (high SLC20A1 group) was 8.34±2.64 and SLC20A1 7.589 0.006 4.16±2.39 cm3 in patients with low SLC20A1 H‑scores (low Higha 12 14 SLC20A1 group) (P<0.05), and the rate of tumor recurrence was Low 3 23 8/26 (30.8%) compared with 3/26 (11.5%) in these two groups, Wif1 9.369 0.002 respectively. Patients with high Wif1 H‑scores (high Wif1 High 2 24 group) had less tumor volume (3.14±2.35 vs. 9.36±3.64 cm3) Low 13 13 than those with low Wif1 H‑scores (low Wif1 group) (P<0.05) and lower recurrence rates compared with patients in the low β‑catenin 4.591 0.032 Wif1 group, data not shown (2/26 vs. 9/26, χ2=5.65, P=0.017). High 11 15 The H‑scores of SLC20A1 were negatively associated with Low 4 22 the H‑scores of β‑catenin in somatotroph adenomas, with a a correlation coefficient of 0.366 (Fig. 2A). The correlation coef- High, positive percentage of SLC20A1, Wif1 or β‑catenin is ≥50%; ficient was ‑0.424 between Wif1 and SLC20A1 (Fig. 2B). PFS low, positive percentage of SLC20A1, Wif1 or β‑catenin <50%. GH, growth hormone; SLC20A1, solute carrier family 20 member 1; associated with low SLC20A1 H‑scores was longer than that Wif1, Wnt inhibitory factor 1. associated with high SLC20A1 H‑scores (P=0.033, Fig. 3A), while the PFS in the high Wif1 group was longer than that in the low Wif1 group (P=0.043), as shown in Fig. 3B.

Effects of RNAi‑SLC20A1 on cell growth, GH release in GH3 cells. The GH levels in the Sh‑B and Sh‑C groups and migration of GH3 cells. The efficiency of four shRNA were 34.7±10.4 and 54.6±14.4% of that in control GH3 cells, segments (Sh‑A/B/C/D‑SLC20A1) was measured in GH3 respectively (P<0.05; Fig. 5B). The transmembrane invasion cells. RT‑qPCR demonstrated that the relative percentages assay revealed that RNAi‑SLC20A1 significantly suppressed of SLC20A1 mRNAs were 32.3±10.4, 8.2±5.7, 17.7±8.4 and cell invasion in the Sh‑B and Sh‑C groups (Fig. 5C). 64.9±22.4%, respectively, after 72 h of transfection (Fig. 4A). RT‑qPCR revealed that SLC20A1 silencing reduced Western blotting also confirmed that the interference effi- the mRNA levels to 0.105±0.114 and 0.233±0.141‑fold for ciency of segment Sh‑B was 83.4% and that of Sh‑C was 76.4% N‑cadherin (N‑CAD), 0.216±0.185 and 0.346±0.172‑fold compared with the SLC20A1 protein levels in GH3 cells after for vimentin, 0.294±0.155 and 0.387±0.212‑fold for vascular 72 h of transfection (Fig. 4B). Cell proliferation experiments endothelial growth factor (VEGF), and 0.344±0.165 and revealed that the cell viability of the Sh‑B group was 76.3±4.5, 0.547±0.282‑fold for matrix metalloproteinase‑2 compared 65.7±3.7 and 53.1±3.2% of control GH3 cells after 24, 48 and with the control group. There was no significant difference 72 h of transfection, respectively, and that of the Sh‑C group in the expression levels of E‑cadherin, matrix metallopro- was 86.4±5.7, 75.6±4.4 and 67.5±3.8% of the control group, teinase‑9 and Snail. In additional, no significant difference respectively, after the same time (P<0.05; Fig. 5A). ELISA was observed between the control group and scramble was used to measure the effect of SLC20A1 on GH secretion group (Fig. 5D). 3280 Molecular Medicine REPORTS 20: 3276-3284, 2019

Figure 1. Expression of SLC20A1, Wif1, β‑catenin and Ki67, assessed by tissue microarray in 52 somatotroph adenomas. (A) Immunohistochemistry analysis. Scale bar, 60 µm. (B) Western blotting showed higher Wif1 and sFRP4 protein levels and lower β‑catenin level in non‑invasive adenomas compared with invasive adenomas. *P<0.05, **P<0.01 vs. respective non‑invasive group. SLC20A1, solute carrier family 20 member 1; Wif1, Wnt inhibitory factor 1; sFRP4, secreted frizzled‑related protein 4.

Figure 2. Correlation of SLC20A1, β‑catenin and Wif1. Scatter diagrams demonstrating the correlation between (A) SLC20A1 and β‑catenin and (B) between SLC20A1 and Wif1 following H‑score analysis of tissue samples from somatotroph adenomas. SLC20A1, solute carrier family 20 member 1; Wif1, Wnt inhibitory factor 1.

SLC20A1 silencing inhibits the activation of the Wif1/β‑ Discussion catenin signaling pathway. Aberrant expression of the Wif1 and sFRP genes has been reported to be associated with Somatotroph adenomas are pituitary tumors that secrete GH the malignant and invasive characteristics of certain human and IGF‑1, resulting in chronic systemic disease with compli- tumors, including somatotroph adenomas (17,18). RT‑qPCR cations and increased mortality when not adequately treated. revealed that the mRNA levels of Wif1 in the Sh‑B and Sh‑C To date, no novel recurrent mutated genes have been identified groups were 4.3±1.2 and 3.6±0.7‑fold, respectively, of those in sporadic adenomas except for GNAS (19,20). By allowing in the control GH3 cells, while the mRNA levels of sFRP4 the cells to move more quickly out of the G0/G1 phase of the cell were 5.8±1.3 and 4.6±0.9‑fold of those in the control GH3 cells cycle, overexpression of SLC20A1 could confer faster adhe- (Fig. 6). Western blotting detected that the Sh‑B and Sh‑C sion and migration to the NIH3T3 cell line (10). The present segments of SLC20A1 could evidently increase the levels of study evaluated the levels of SLC20A1 and the Wnt/β‑catenin Wif1 and sFRP4, and significant decrease the β‑catenin level signaling pathway, and their role in the tumorigenesis and in the Sh‑B group (Fig. 6). clinical features of 52 somatotroph adenomas. LI et al: SLC20A1 IN SOMATOTROPH ADENOMAS 3281

Figure 3. Average PFS in 52 patients with somatotroph adenomas. (A) Average PFS according to SLC20A1 levels. (B) Average PFS according to Wif1 levels. Blue line, high‑level group; green line, low‑level group. PFS, progression‑free survival; SLC20A1, solute carrier family 20 member 1; Wif1, Wnt inhibitory factor 1.

Figure 4. Efficiency of four shRNA segments in GH3 cells. (A) Reverse transcription‑quantitative PCR data showing that the mRNA levels of SLC20A1 in the Sh‑A‑D groups were reduced to 32.3±10.4, 8.2±5.7, 17.7±8.4 and 64.9±22.4%, respectively, of the mRNA levels in control GH3 cells after 72 h of transfection. (B) Western blot analysis showing that the SLC20A1 level in the Sh‑B and Sh‑C groups reduced to 16.6±4.7 and 23.6±5.3%, respectively, of the control group. n=3. *P<0.05, **P<0.01 vs. respective control. sh, short hairpin; SLC20A1, solute carrier family 20 member 1.

Compared with other subtypes of pituitary adenomas, PFS in patients with low SLC20A1 and/or high Wif1. There somatotroph adenomas are associated with higher morbidity are few cell lines used for the study of pituitary adenomas. The and mortality. Main risk factors include histological subtype MMQ cell line (25) is used for prolactinomas, the GH3 cell and tumor invasion as well as the presence of molecular line (26) is used for somatotroph adenomas and the Att20 cell biomarkers, including Ki67, p53 and IGF‑1 (21,22). Epidermal line (27) is used for corticotroph adenomas. GT1‑1 cells (28) growth factor‑like domain multiple 7 protein and sparsely mainly secrete luteinizing hormone and low levels of GH. granulated adenomas may serve as useful biomarkers of In vitro, the Wnt/β‑catenin signaling pathway may be involved somatotroph adenoma invasion and prognosis (23). However, in the cell proliferation, migration and GH secretion of GH3 several predictive biomarkers have failed in clinical research cells via overexpression of SLC20A1. trials of neuroendocrine tumors (24). There is an urgent unmet SLC20A1 is transcribed in response to the activation of need to incorporate novel biomarker candidates in order to nuclear factor‑κB (NF‑κB) during cancer cell progression (29). obtain robust prospective validation. The present IHC results Furthermore, SLC20A1 inhibits tumor necrosis factor‑induced indicated that high expression of SLC20A1 may promote the apoptosis through the main antiapoptotic signals triggered by malignant potential of somatotroph adenomas. High levels of the mitogen‑activated protein kinase‑Jun N‑terminal kinase expression of SLC20A1 and low levels of expression of Wif1 signaling pathway (30,31). In fact, NF‑κB and β‑catenin are were observed in somatotroph adenomas, as well as a longer constitutively activated by upstream serine/threonine kinases 3282 Molecular Medicine REPORTS 20: 3276-3284, 2019

Figure 5. RNAi‑SLC20A1 suppresses cell proliferation, GH secretion and migration of GH3 cells. (A) Effect on cell viability, as shown by MTS assay. (B) ELISA revealed that RNAi‑SLC20A1 could reduce GH secretion. (C) The effect on invasion, as shown by the transwell experiment. Magnification, x200. (D) The mRNA level of genes associated with migration and invasion upon RNAi‑SLC20A1 in GH3 cells. n=3. *P<0.05, **P<0.01 vs. respective control. RNAi, RNA interference; SLC20A1, solute carrier family 20 member 1; GH, growth hormone; sh, short hairpin; E‑CAD, E‑cadherin; N‑CAD, N‑cadherin; MMP, matrix metalloproteinase; VEGF, vascular endothelial growth factor.

Figure 6. RNAi‑PIT inhibits the Wnt/β‑catenin pathway. (A) Reverse transcription‑quantitative PCR analysis revealed significant changes in Wif1, β‑catenin and sFRP4 mRNA levels in GH3 cells. (B) Western blot analysis revealed a significant change in Wif1, β‑catenin and sFRP4 proteins in the Sh‑B and Sh‑C groups compared with the control group. The protein levels of Wif1 in the Sh‑B and Sh‑C groups were 4.8±0.52 and 4.2±0.42‑fold greater, respectively, than those in the control GH3 cells, while the protein levels of sFRP4 in the Sh‑B and Sh‑C groups were 3.8±0.43 and 3.5±0.37‑fold greater, respectively, than those in the control GH3 cells. n=3. *P<0.05, **P<0.01 vs. respective control. RNAi, RNA interference; SLC20A1, solute carrier family 20 member 1; Wif1, Wnt inhibitory factor 1; sFRP4, secreted frizzled‑related protein 4; sh, short hairpin. implicated in malignant transformation, including apoptosis aggressiveness in 212 patients with NFPAs (33). In a previous evasion (32). There was a significant positive correlation reported study, RT‑qPCR confirmed the lower mRNA levels of Wif1 between the H‑scores of β‑catenin and c‑myc expression and and sFRP2/4 in all pituitary adenoma subtypes, including LI et al: SLC20A1 IN SOMATOTROPH ADENOMAS 3283 somatotroph adenomas, compared with normal pituitary speci- Authors' contributions mens (34). In the present study, SLC20A1 and β‑catenin were increased in the invasive somatotroph adenomas compared with JL contributed to the writing and the IHC experiment. WD the non‑invasive somatotroph adenomas or normal pituitary contributed to the IHC experiment. ZL performed the WB glands. A positive correlation between SLC20A1 and β‑catenin, experiment and follow up. HW performed the cell culture. HG and a negative correlation between SLC20A1 and Wif1, was was involved in the cell function experiments and revising the identified by IHC. Follow‑up data also revealed the potential manuscript. YZ designed the experiment and performed the of SLC20A1 and Wif1 as predictive factors for somatotroph statistical analysis. adenoma recurrence due to high recurrence rate and shorter time to recurrence in patients with high SLC20A1 or low Wif1. Ethics approval and consent to participate The Wnt/β‑catenin signaling pathway plays an important role in the adhesion, invasion and metastasis of pituitary tumors, The protocols were approved by the Internal Review Board of as well as development of these tumors and tumorigenesis (35). Beijing Tiantan Hospital affiliated to Capital Medical University The activation of this pathway is an important factor for VEGF and the study was conducted according to the principles of the production during the process of angiogenesis. Wif1 overexpres- Declaration of Helsinki (no. KY‑2013‑02). Informed consent sion plays an important role in the suppression of cell migration, was obtained from all patients. and Wif1 knockdown enhances the migratory potential of glio- blastomas through Wnt family member 5A activation in vitro and Patient consent for publication in an orthotopic brain tumor model (36,37). In vitro, silencing of SLC20A1 inhibits the proliferation and migration of GH3 cells. Not applicable. Western blotting revealed that SLC20A1 could reduce the Wif1 and sFRP4 protein levels due to higher Wif1 and sFRP4 in the Competing interests Sh‑B or Sh‑C groups compared with the control or scramble groups. Epithelial‑mesenchymal transition (EMT) markers such The authors declare that they have no competing interests. as N‑cad, vimentin and Snail are important indicators of the appearance of cystic lesions, tumor progression, bone destruc- References tion and endocrine functions in somatotroph adenomas (38). The present data revealed that SLC20A1 evidently regulated 1. AlDallal S: Acromegaly: A challenging condition to diagnose. Int J Gen Med 11: 337‑343, 2018. the mRNA levels of N‑CAD, VEGF and vimentin, which are 2. Inoshita N and Nishioka H: The 2017 WHO classification of pitu- associated with cell invasion and migration, after 72 h of RNAi. itary adenoma: Overview and comments. Brain Tumor Pathol 35: Taken together, these data suggest a novel mechanism for the 51‑56, 2018. 3. Abreu A, Tovar AP, Castellanos R, Valenzuela A, Giraldo CM, role of SLC20A1 in cell proliferation through the Wnt/β‑catenin Pinedo AC, Guerrero DP, Barrera CA, Franco HI, Ribeiro- signaling pathway in GH3 cells. Oliveira A Jr, et al: Challenges in the diagnosis and management In conclusion, the present study demonstrated that a high of acromegaly: A focus on comorbidities. Pituitary 19: 448‑457, 2016. SLC20A1 level is positively associated with the tumor size, inva- 4. Kempf J, Schmitz A, Meier A, Delfs N, Mueller B, Fandino J, sive behavior and tumor recurrence of somatotroph adenomas. Schuetz P and Berkmann S: Adenoma size and postoperative Furthermore, SLC20A1 may be associated with activation of IGF‑1 levels predict surgical outcomes in acromegaly patients: β in vitro Results of the Swiss Pituitary Registry (SwissPit). Swiss Med the Wnt/ ‑catenin signaling pathway. Furthermore, Wkly 148: w14653, 2018. experiments demonstrated that silencing SLC20A1 inhibited 5. Välimäki N, Demir H, Pitkänen E, Kaasinen E, Karppinen A, cell proliferation and invasion through inhibition of genes Kivipelto L, Schalin‑Jäntti C, Aaltonen LA and Karhu A: associated with EMT. However, the present study had certain Whole‑genome sequencing of growth hormone (GH)‑secreting pituitary adenomas. J Clin Endocrinol Metab 100: 3918‑3927, limitations. Due to the small sample size and possible selection 2015. bias in the present study, it is necessary to confirm and validate 6. Picard C, Silvy M, Gerard C, Buffat C, Lavaque E, the conclusions in further studies with larger sample sizes. Figarella‑Branger D, Dufour H, Gabert J, Beckers A, Brue T, et al: Gs alpha overexpression and loss of Gs alpha imprinting in human somatotroph adenomas: Association with tumor size Acknowledgements and response to pharmacologic treatment. Int J Cancer 121: 1245‑1252, 2007. 7. Forster IC, Hernando N, Biber J and Murer H: Phosphate trans- Not applicable. porters of the SLC20 and SLC34 families. Mol Aspects Med 34: 386‑395, 2013. Funding 8. Ravera S, Virkki LV, Murer H and Forster IC: Deciphering PiT transport kinetics and substrate specificity using electrophysi- ology and fux measurements. Am J Physiol Cell Physiol 293: The present study was supported by Beijing Natural Science C606‑C620, 2007. Foundation of China (grant no. 7162035), Beijing High Level 9. Byskov K, Jensen N, Kongsfelt IB, Wielsøe M, Pedersen LE, Haldrup C and Pedersen L: Regulation of cell proliferation and Program (grant no. 2015‑3‑040) and the National Natural cell density by the inorganic phosphate transporter PiT‑1. Cell Science Foundation of China (grant no. 81601205). Division 7: 7, 2012. 10. Kongsfelt IB, Byskov K, Pedersen LE and Pedersen L: High levels of the type III inorganic phosphate transporter PiT‑1 (SLC20A1) Availability of data and materials can confer faster cell adhesion. Exp Cell Res 326: 57‑67, 2014. 11. Sato K and Akimoto K: Expression levels of KMT2C and The datasets used and/or analyzed during the current study are SLC20A1 identified by information‑theoretical analysis are powerful prognostic biomarkers in ER‑positive breast cancer. available from the corresponding author on reasonable request. Clin Breast Cancer 17: e135‑e142, 2017. 3284 Molecular Medicine REPORTS 20: 3276-3284, 2019

12. Clevers H and Nusse R: Wnt/β‑catenin signaling and disease. 27. Bergeron F, Sirois F and Mbikay M: ACTH secretion by mouse Cell 149: 1192‑1205, 2012. corticotroph AtT20 cells is negatively modulated by the intracel- 13. Song W, Qian L, Jing G, Jie F, Xiaosong S, Chunhui L, lular level of 7B2. FEBS Lett 512: 259‑262, 2002. Yangfang L, Guilin L, Gao H and Yazhuo Z: Aberrant expres- 28. Peng H, Fan J, Wu J, Lang J, Wang J, Liu H, Zhao S and Liao J: sion of the sFRP and WIF1 genes in invasive non‑functioning Silencing of HEPN1 is responsible for the aggressive biological pituitary adenomas. Mol Cell Endocrinol 474: 168‑175, 2018. behavior of pituitary somatotroph adenomas. Cell Physiol 14. Ricardo VL, Robert YO, Gunter K and Juan R: WHO classifica- Biochem 31: 379‑388, 2013. tion of tumours of endocrine organ (4th edition). International 29. Karin M: Nuclear factor‑kappaB in cancer development and agency for research on cancere. France, Lyon, pp.19‑23, 2018. progression. Nature 441: 431‑436, 2006. 15. Liu C, Gao H, Cao L, Gui S, Liu Q, Li C, Li D, Gong L and 30. Salaün C, Leroy C, Rousseau A, Boitez V, Beck L and Zhang Y: The role of FSCN1 in migration and invasion of pitu- Friedlander G: Identification of a novel transport‑independent itary adenomas. Mol Cell Endocrinol 419: 217‑224, 2016. function of PiT‑1/SLC20A1 in the regulation of TNF‑induced 16. Livak KJ and Schmittgen TD: Analysis of relative gene expres- apoptosis. J Biol Chem 285: 34408‑34418, 2010. sion data using real‑time quantitative PCR and the 2(‑Delta Delta 31. Wajant H, Pfizenmaier K and Scheurich P: Tumor necrosis factor C(T)) method. Methods 25: 402‑408, 2001. signaling. Cell Death Differ 10: 45‑65, 2003. 17. Pez F, Lopez A, Kim M, Wands JR, Caron de Fromentel C and 32. Kavitha K, Kowshik J, Kishore TK, Baba AB and Nagini S: Merle P: Wnt signaling and hepatocarcinogenesis: Molecular Astaxanthin inhibits NF‑κB and Wnt/β‑catenin signaling path- targets for the development of innovative anticancer drugs. ways via inactivation of Erk/MAPK and PI3K/Akt to induce J Hepatol 59: 1107‑1117, 2013. intrinsic apoptosis in a hamster model of oral cancer. Biochim 18. Ueland T, Olarescu NC, Jørgensen AP, Otterdal K, Aukrust P, Biophys Acta 1830: 4433‑4444, 2013. Godang K, Lekva T and Bollerslev J: Increased serum and bone 33. Liu C, Wu Y, Yu S, Bai J, Li C, Wu D and Zhang Y: Increased matrix levels of the secreted Wnt antagonist DKK‑1 in patients β‑catenin and c‑myc expression predict aggressive growth of with growth hormone deficiency in response to growth hormone non‑functioning pituitary adenomas: An assessment using a treatment. J Clin Endocrinol Metab 100: 736‑743, 2015. tissue microarray‑based approach. Mol Med Rep 5: 1793‑1799, 19. Hage M, Viengchareun S, Brunet E, Villa C, Pineau D, 2017. Bouligand J, Teglas JP, Adam C, Parker F, Lombès M, et al: 34. Elston MS, Gill AJ, Conaglen JV, Clarkson A, Shaw JM, Genomic alterations and complex subclonal architecture in Law AJ, Cook RJ, Little NS, Clifton‑Bligh RJ, Robinson BG and sporadic GH‑secreting pituitary adenomas. J Clin Endocrinol McDonald KL: Wnt pathway inhibitors are strongly down‑regu- Metab 103: 1929‑1939, 2018. lated in pituitary tumors. Endocrinology 149: 1235‑1242, 2008. 20. Caimari F and Korbonits M: Novel genetic causes of pituitary 35. Chambers TJ, Giles A, Brabant G and Davis JR: Wnt signal- adenomas. Clin Cancer Res 22: 5030‑5042, 2016. ling in pituitary development and tumorigenesis. Endocr Relat 21. Mete O and Lopes MB: Overview of the 2017 WHO Classification Cancer 20: R101‑R111, 2013. of pituitary tumors. Endocr Pathol 28: 228‑243, 2017. 36. Vassallo I, Zinn P, Lai M, Rajakannu P, Hamou MF and Hegi ME: 22. Park HH, Kim EH, Ku CR, Lee EJ and Kim SH: Outcomes WIF1 re‑expression in glioblastoma inhibits migration through of aggressive surgical resection in growth hormone‑secreting attenuation of non‑canonical WNT signaling by downregulating pituitary adenomas with cavernous sinus invasion. World the lncRNA MALAT1. Oncogene 35: 12‑21, 2016. Neurosurg 117: e280‑e289, 2018. 37. Wang Y, Sang A, Zhu M, Zhang G, Guan H, Ji M and Chen H: 23. Wang J, Liu Q, Gao H, Wan D, Li C, Li Z and Zhang Y: Tissue factor induces VEGF expression via activation of the EGFL7 participates in regulating biological behavior of growth Wnt/β‑catenin signaling pathway in ARPE‑19 cells. Mol Vis 22: hormone‑secreting pituitary adenomas via Notch2/DLL3 886‑897, 2016. signaling pathway. Tumour Biol 39: 1010428317706203, 2017. 38. Shan XS, Liu Q, Li ZY, Li CZ, Gao H and Zhang YZ: 24. Barriuso J, Custodio A, Afonso R, Alonso V, Astudillo A, Epithelial‑mesenchymal transition induced by SMAD4 activa- Capdevila J, García‑Carbonero R, Grande E, Jimenez‑Fonseca P, tion in invasive growth hormone‑secreting adenomas. Open Marazuela M, et al: Prognostic and predictive biomarkers for Chem 16: 571‑582, 2018. somatostatin analogs, peptide receptor radionuclide therapy and serotonin pathway targets in neuroendocrine tumours. Cancer Treat Rev 70: 209‑222, 2018. 25. Su Z, Cai L, Lu J, Li C, Gui S, Liu C, Wang C, Li Q, Zhuge Q and This work is licensed under a Creative Commons Zhang Y: Global expression profile of tumor stem‑like cells isolated Attribution-NonCommercial-NoDerivatives 4.0 from MMQ rat prolactinoma cell. Cancer Cell Int 17: 15, 2017. International (CC BY-NC-ND 4.0) License. 26. Occhi G, Losa M, Albiger N, Trivellin G, Regazzo D, Scanarini M, Monteserin‑Garcia JL, Fröhlich B, Ferasin S, Terreni MR, et al: The glucose‑dependent insulinotropic polypeptide receptor is overexpressed amongst GNAS1 mutation‑negative somatotro- pinomas and drives growth hormone (GH)‑promoter activity in GH3 cells. J Neuroendocrin 23: 641‑649, 2011.